INDEX

Explanations

phrases related to shutting things down or turning things off

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enegger

-0.72

ozy

-0.71

arya

-0.69

omething

-0.68

lihood

-0.68

 IMAGES

-0.67

xual

-0.67

ãĥīãĥ©ãĤ´ãĥ³

-0.66

 Plenty

-0.65

intent

-0.63

POSITIVE LOGITS

tered

1.35

tering

1.18

downs

1.01

down

0.97

 down

0.96

shut

0.90

 doors

0.88

 shut

0.83

 downs

0.83

ters

0.81

Activations Density 0.556%

phrases related to shutting things down or turning things off

No Comments

No Known Activations